GPSto80to90%( 27 ), while maintaining high
spatial accuracy. This is particularly important
for small and nocturnal animals such as bats,
for which solar-recharging technologies are
impractical. Consequently, ATLAS opens new
opportunities for investigating the movement
ecology of bats, which account for ~25% of
known mammal species and of which at least
25% are threatened ( 28 ). More generally,
ATLAS suggests a promising new avenue for
animal movement and navigation research.
Recent progress in neuroscience has yielded
important insights into the cellular-level pro-
cesses underlying animal behavior. Such lab
experiments are vital but might not reveal
how animals behave or utilize their cognitive
potential in the wild ( 29 , 30 ). Using multiple
approaches, analyses, and reasonings imple-
mented on our high-resolution dataset, we
have presented here several lines of evidence
for the existence of a cognitive map in wild
animals under natural settings, crucially com-
plementing lab-based evidence ( 7 , 9 , 10 ). We
caution, however, that although all these re-
sults are consistent with our interpretation,
two main caveats [highlighted in ( 18 )] should
be carefully considered. First, lacking infor-
mation on bats’tracks before we trapped and
tagged them, shortcuts or unrecorded routes
cannot be considered as strictly novel. Yet, they
maintain the same properties (straight ab initio
goal-directed flights) regardless of when the bats
first traversed them: more than a month after
tagging on average, and also after 5 months,
across fruiting seasons, of continuous high-
resolution tracking (Fig. 3D). Furthermore, 3-
to 8-month-old juveniles also exhibited straight
ab initio goal-directed shortcuts that were indis-
tinguishable from those of the adults, and fruit
bats in our study rarely (0.4%) fly together even
for a short part of their commuting flights. We
thus propose that a cognitive map is already
present at an early stage and is likely acquired
within the first 3 months of life. Having high-
resolution lifetime (from birth to death) move-
ment tracks of wild bats of various ages could
greatly complement this study, to elaborate at
which early stage of life a cognitive map is
acquired (for juveniles) and for how long the
spatial memory lasts across seasons and years
(for adults). Second, tracked bats might still
have used alternative (simpler) navigation
mechanisms such as piloting and beaconing,
as animals navigating in real-life environments
likely use more than one navigation mecha-
nism ( 30 ). Consequently, we do not claim that
bats foraging within their familiar home range
exclusively rely on a cognitive map, but rather
that our results (i) are consistently compatible
with multiple features expected from cognitive
map–based navigation; (ii) are strictly incom-
patible with expectations of random search and
path integration; and (iii) do not lend support
to alternative navigation mechanisms that
we examined, such as piloting and beacon-
ing, although additional field experiments
are needed to divulge whether, and to what
extent, wild bats also use such mechanisms.
We have previously employed such field ex-
periments, revealing a key role for visual land-
marks in large-scale homing of fruit bats ( 11 ).
We thus hypothesize that a few visual land-
marks might be sufficient for wild bats to
anchor the cognitive map of their entire (rela-
tively large) home range.
We suggest that such complex foraging na-
vigation, considered to reflect high cognitive
ability, may have evolved under selective forces
similar to those proposed for long-lived, forest-
dwelling frugivorous primates ( 31 ). Egyptian
fruit bats feed on spatially patchy yet tempo-
rally predictable long-lived resources (fruit
trees), have extreme longevity (25 years and
more) relative to body size ( 32 , 33 ), and are
social central-place foragers that return home
to a cave at the end of the night rather than
roost on trees within their foraging range. We
suggest that these features may have selected
for extended spatial memory and high forag-
ing navigation performance of this and perhaps
other similar species as well. It is interesting
that Pteropodid bats have the largest relative
brain size among bat families, surpassing even
basal primates ( 34 )—reminiscent of frugivorous
primates that tend to have larger relative brain
sizes than mostly folivorous primates ( 31 ). A gen-
uine integration between lab-based neurobiol-
ogy, experimental animal cognition research,
and field-based movement ecology ( 30 , 35 , 36 )
could open new opportunities to further unravel
the components of spatial cognition. Such inte-
gration may also help in understanding other
long-lasting key questions on the neural, behav-
ioral, and ecological mechanisms underlying
cognitive performance of wild animals in their
natural environments.REFERENCES AND NOTES- J. Wieneret al., inAnimal Thinking: Contemporary Issues in
Comparative Cognition,R. Menzel,
J. Fischer, Eds. (MIT Press, Cambridge, MA, 2011). - H. G. Wallraff,Animal Navigation: Pigeon Homing as a Paradigm
(Springer, Berlin, 2005). - G. Kramer,J. Ornithol. 94 , 201–219 (1953).
- E. C. Tolman,Psychol. Rev. 55 , 189–208 (1948).
- J. O’Keefe, L. Nadel,The Hippocampus as a Cognitive Map
(Clarendon, 1978). - C. R. Gallistel,The Organization of Learning(MIT Press,
1990). - M. Geva-Sagiv, L. Las, Y. Yovel, N. Ulanovsky,Nat. Rev.
Neurosci. 16 , 94–108 (2015). - E. I. Moser, E. Kropff, M.-B. Moser,Annu. Rev. Neurosci. 31 ,
69 – 89 (2008). - A. Finkelsteinet al.,Nature 517 , 159–164 (2015).
- A. Sarel, A. Finkelstein, L. Las, N. Ulanovsky,Science 355 ,
176 – 180 (2017). - A. Tsoaret al.,Proc. Natl. Acad. Sci. U.S.A. 108 , E718–E724
(2011). - S. Greif, I. Borissov, Y. Yovel, R. A. Holland,Nat. Commun. 5 ,
4488 (2014). - A. Weller Weiseret al., inProceedings of the 2016 ACM/IEEE
International Conference on Information Processing in Sensor
Networks (IPSN), Vienna, Austria, 12 to 14 April 2016, pp. 1–12.
14. S. Toledo, O. Kishon, Y. Orchan, A. Shohat, R. Nathan,
Lessons and experiences from the design, implementation, and
deployment of a wildlife tracking system, inProceedings of the
2016 IEEE International Conference on Software Science,
Technology and Engineering (SWSTE), pp. 51–60.
15. Materials and methods are available as supplementary
materials.
16. P.A.Zollner,S.L.Lima,Ecology 80 , 1019– 1030
(1999).
17. E. R. Heithaus, T. H. Fleming,Ecol. Monogr. 48 , 127– 143
(1978).
18. A. T. Bennett,J. Exp. Biol. 199 , 219–224 (1996).
19. F. Takens,“Detecting strange attractors in turbulence”in
Dynamical Systems and Turbulence, Warwick 1980(Springer,
1981), vol. 898, pp. 366–381.
20. U. Nehmzow,Scientific Methods in Mobile Robotics:
Quantitative Analysis of Agent Behaviour(Springer, London,
2006).
21. A. Hastings, C. L. Hom, S. Ellner, P. Turchin, H. C. J. Godfray,
Annu. Rev. Ecol. Syst. 24 ,1–33 (1993).
22. J. E. Skinner,Nat. Biotechnol. 12 , 596–600 (1994).
23. M. T. Rosenstein, J. J. Collins, C. J. De Luca,Physica D 65 ,
117 – 134 (1993).
24. I. Schiffner, J. Baumeister, R. Wiltschko,J. Theor. Biol. 291 ,
42 – 46 (2011).
25. E. Normand, C. Boesch,Anim. Behav. 77 , 1195– 1201
(2009).
26. A. Gagliardo,J. Exp. Biol. 216 , 2165–2171 (2013).
27. R. Kays, M. C. Crofoot, W. Jetz, M. Wikelski,Science 348 ,
aaa2478 (2015).
28. K. E. Jones, A. Purvis, J. L. Gittleman,Am. Nat. 161 , 601– 614
(2003).
29.C.H.Janson,R.Byrne,Anim. Cogn. 10 , 357– 367
(2007).
30. L. F. Jacobs, R. Menzel,Mov. Ecol. 2 ,3–3 (2014).
31. K. Milton,Am. Anthropol. 83 , 534–548 (1981).
32. G. Neuweiler,The Biology of Bats(Oxford Univ. Press,
New York, 2000).
33. G. G. Kwiecinski, T. A. Griffiths,Mamm. Species 611 ,1– 9
(1999).
34. P. Pirlot, H. Stephan, Encephalization in Chiroptera.Can. J. Zool.
48 , 433 (1970).
35. R. Nathanet al.,Proc. Natl. Acad. Sci. U.S.A. 105 , 19052– 19059
(2008).
36. J. Morand-Ferron, E. F. Cole, J. L. Quinn,Biol. Rev. Camb.
Philos. Soc. 91 , 367–389 (2016).
37. D. Shohami, R. Nathan, Cognitive map-based navigation
in wild bats revealed by a new high-throughput tracking
system, Dryad (2020); https://doi.org/10.5061/dryad.
g4f4qrfn2.
ACKNOWLEDGMENTS
For valuable fieldwork assistance, we thank R. Lotan, A. Levi,
S. Margalit, and other Movement Ecology Lab and the Minerva Center
for Movement Ecology members. We thank N. Ulanovsky and
A. Bennett for helpful comments on an earlier version, A. Ben-Nun
for help with GIS analysis, and Y. Yovel for discussions on this
topic;Funding:ATLAS development, maintenance, and studies
have been supported by the Minerva Center for Movement Ecology,
the Minerva Foundation, and ISF grant ISF-965/15; bat research
in the movement ecology lab was supported also by grants from
ISF-1316/05, ISF-1259/09, and GIF 1316/15. We also acknowledge
support from Adelina and Massimo Della Pergola Chair of Life
Sciences to R.N. and the Israel President Scholarship to D.S.
Author contributions:D.S. and R.N. conceived the study; R.N.
conceived and S.T. developed the ATLAS system, with the help of
Y.O. and Y.B.; D.S. and E.L. carried out fieldwork; D.S. analyzed the
data; I.S. conceived and performed the time-lag embedding
analysis; D.S. and R.N. wrote the manuscript with input from all
other coauthors;Competing interests:The authors declare no
competing interests.Data and materials availability:All ATLAS
localization data used in the analysis are available on Dryad ( 37 ).SUPPLEMENTARY MATERIALS
science.sciencemag.org/content/369/6500/188/suppl/DC1
Materials and Methods
Figs. S1 to S4
Tables S1 to S4
References ( 38 – 42 )6 May 2019; accepted 29 May 2020
10.1126/science.aax6904SCIENCEsciencemag.org 10 JULY 2020•VOL 369 ISSUE 6500 193
RESEARCH | REPORTS